John Mandrola, MD


November 12, 2017

On the first day of the American Heart Association (AHA) 2017 Scientific Sessions , I've made a habit of attending the Samuel A Levine Young Investigators sessions.

This year's session did not disappoint. The judges will have tough choices. Two of the five talks even had simultaneous publications.

Since I can't cover all the talks, I chose the one that impressed me most. Preventive cardiology won out. Really.

Dr Amit V Khera (Massachusetts General Hospital, Boston) presented his team's work using a whole-genome sequence to predict early-onset MI.[1] Don't be scared by the words "genome." Keep reading, I promise, this is neat stuff.

Here are the key points:

Everyone knows that most diseases, like early MI, come from an interplay of lifestyle and genes (polygenetic). Less common are high-risk single-gene (monogenic) mutations, such as familial hypercholesterolemia (heart disease) and BRCA (breast cancer).

In the past, research focused on single-gene causes of disease. Recent tech has made possible the ability to look at the entire DNA in aggregate. The grand idea of polygenetic scores is to use the amalgam of base pairs we get from mom and dad to predict future disease.

Khera's study took genome-sequencing data from nearly 2400 people who had had MI before age 55. They then compared these with a matched control group of around 4200 patients.

About 2% of the MI group had the single-gene variant of FH. The odds ratio for MI in the FH group was 3.2. That's not unexpected, as their LDL-C levels were 53 points higher than controls.

The key question for Khera and colleagues was: How would an aggregate (polygenetic) gene-risk score compare with a known single-gene mutation like FH?

Previous work with polygenetic risk scores have included only the top 50 to 100 known variants associated with heart disease. For this study, the investigators went further—way further.

They weighted 6.6 million variants of base pairs to come up with a polygenetic score. That's right: 6.6 million base pairs! Senior investigator Dr Sekar Kathiresan (Massachusetts General Hospital) told me after the presentation using this many variants was actually not that difficult.

Because gene scores can be arbitrary on how to define what exactly is "high" risk, Khera's team studied the highest 20%, then 10%, and then 5% of gene risk scores. Picture the gene scores as a bell curve; they progressively studied narrower parts of the right-hand tail of risk.

The odds ratio for early MI in the highest 5% of risk was 3.6. In other words, about one in 20 individuals had nearly a four-times-greater risk of MI than the 95% of the population who did not have this score. That's a higher risk than patients with FH—a group we offer early preventive therapies.

How is this different from a simple family history? It's a lot better. Only 39% of the individuals in this 5% high polygenetic risk reported a positive family history. Translation: the majority of those at nearly fourfold higher genetic risk did not have a positive family history.

Khera's group tested their score in a larger population using the UK Biobank (n=500,000). Their score held up. Get this, they found those in the top 0.25% of gene scores had a sixfold greater rate of MI before age 55. That's about the same risk BRCA mutations confer for breast cancer.

Perhaps you can see why this is so vital: Heart disease is preventable—despite high genetic risk. Khera is primary author of a New England Journal of Medicine paper showing that those individuals with a high genetic risk who maintained a healthy lifestyle almost halved their MI risk compared with high gene score folks who did not.[2]

Why do we need gene scores when we have calcium scores? The reason gene scores may be better than phenotype tests, such as a calcium score, is that you can know your genetic risk from childbirth. That's key because atherosclerosis risk turns on long-term exposure to lifestyle factors or high LDL-C.

Finally, and this is what's really cool: Kathiresan told me these sorts of gene scores are in the works for other important diseases, such as obesity and breast cancer.

Using DNA to predict cancer risk has implications for screening. One reason screening for major diseases does not improve overall mortality[3] may be that harm from the many false positives create a drag on net benefits of early detection. But screening and/or intervening in those with high genetic risks (think precision) may improve both the public's and an individual's health.

If you like science and its potential, you have to love gene risk scores.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.